One of greatest challenges in oncology is eradicating cancer once it has metastasized. Cancer immunotherapy (IMT) with checkpoint blockade drugs like ?-PD1 has shown remarkable sustained complete responses in a subset of cancer patients. But even within sensitive tumor types such as melanoma, only around 25% of patients respond. Many challenges remain to fully capitalize on existing and novel drugs that reactivate anti- tumor immunity. There is, therefore, an urgent unmet need for mouse models that better reflect human cancer. Our goal is to utilize the well characterized chemically-induced cutaneous squamous cell carcinoma (cSCC) model to i) identify genetic features of tumors that establish IMT sensitivity, ii) define mechanisms and biomarkers of innate and acquired ?-PD-1 drug resistance iii) define mechanisms of cooperation between ?TGF? and ?PD1 therapy, and causes of failure. We will build on two novel and important observations that we have made: First, that a new ?-panTGF? antibody is as effective as IMT and cooperates with ?PD-1 to double the IMT response rate, eliciting robust tumor rejection and sustained tumor immunity in the cSCC model. Secondly, just as in human cancers, we find that response to IMT is higher in chemically-induced cSCCs with high mutational single nucleotide variant (SNV) loads than those with low SNV load, including a genetically engineered mouse model (GEMM). In Aim 1 we will address how ?TGF? therapy enhances ?PD-1 responses, by investigating drug effects on tumor immune cell numbers and function. We will validate our preliminary finding of Treg involvement in cooperativity between ?PD1 and ?TGF? drugs using CyTO and MIBI analysis. We will also investigate involvement of other immune cell types, particularly myeloid, that may contribute to ?TGF?/?PD-1 drug cooperativity. In Aim 2, we will i dentify tumor cell genetic properties that determine IMT responses, and events driving development of IMT drug resistance, in a panel of independent DMBA/TPA induced syngeneic primary cSCC lines driven by distinct chemically-induced mutations of Kras or Hras. We will investigate effects of a) distinct Ras driver mutations b) SNV loads and neoantigen quality, and c) miRNA and RNA transcriptome profiles, on ?-PD-1 and/or ?TGF? responses, and we will extend our studies into primary chemically induced sSCC, and validate the use of predictive biomarkers found in Aims 1 and 2 in this more realistic and heterogeneous model of human cancer. Finally, in Aim 3, we will validate our preclinical findings of mutational, transcriptomic and/or immune signatures predictive of IMT responses by interrogating pre-treatment and post-treatment clinical HNSCC samples from ?PD-1 responding versus non-responding HNSCC patients under treatment at UCSF. By project completion, we will have validated a translational model for IMT that may be utilized by others for novel IMT agent development, and for mechanistic studies that will provide personalized treatment options to cancer patients.